Hey Guys,
 

my monthly Parity check finished and I got 404 Sync Errors. I am a little worried but also had to smile because of 404. My monthly check runs without correcting errors, so I am currently running the second parity check with corrections on. I have combed through the logs and think that the errors occured on my second parity disk because of this
 

Dec  1 04:00:01 Jarvis kernel: mdcmd (65): check NOCORRECT
Dec  1 04:00:01 Jarvis kernel: 
Dec  1 04:00:01 Jarvis kernel: md: recovery thread: check P Q ...
Dec  1 07:22:00 Jarvis kernel: md: recovery thread: Q incorrect, sector=3908615760
Dec  1 07:22:00 Jarvis kernel: md: recovery thread: Q incorrect, sector=3908615768
Dec  1 07:22:00 Jarvis kernel: md: recovery thread: Q incorrect, sector=3908615776
Dec  1 07:22:00 Jarvis kernel: md: recovery thread: Q incorrect, sector=3908615784
Dec  1 07:22:00 Jarvis kernel: md: recovery thread: Q incorrect, sector=3908615792
Dec  1 07:22:00 Jarvis kernel: md: recovery thread: Q incorrect, sector=3908615800

 

There were more lines but i will leave them out because they are the same but increasing in sector.  
I have also attached the logs. Be aware i went a little overboard on smart tests. I first started an extended test on every disk only afterwards realizing that a short one would suffice. Then stopping every extended and starting a short one on every disk. There were no errors with the disks apart from a few dead sectors that were there from day one(with the parity 2 disks having no errors at all). The array itself also doesn't have any errors and i couldn't find anything in the log pointing me to something. I had a few system hangups in the last months because of VMs but that would have before the monthly test of november but that returned 0 errors. So i am a little stumped to what it could be. 
I also read it could be RAM but I got 96 gigs of ECC memory which hopefully would prevent something like that. 
I thought it could have been me shuffling multiple 100 gigs of data around while the parity check is running but those errors occured while i was sleeping so that couldn't have been it.

I hope I wrote down everything important in here if not I am sorry and happy to give further info. 

 

jarvis-diagnostics-20201201-2207.zip

9 hours ago, Jakobustrop said:

Q incorrect

This means errors were limited to parity2, any unclean shutdowns since last check?

Not that I know of. Last shutdown was 35 days ago and the previous Parity check didn't Return errors.

Run another one without rebooting and post the diags if it finds more errors.

So the second scan finished correcting the 404 Sync errors. No new errors appeared. 

jarvis-diagnostics-20201202-1733.zip

That's never good if there wasn't any unclean shutdown, but if the second check didn't find more errors not much else you can do now, see if future checks find more.

Okay, thank you! Imma keep an eye on it.